(326 days)
The Neoasis® device is intended to reduce noise levels inside an infant incubator in the neonatal intensive care unit (NICU). The attenuation performance of the device is for noises in the NICU with frequencies between 250 Hz to 1,000 Hz.
The Neoasis is comprised of an outside noise sensor, a control unit, two speakers, and a residual noise sensor. All components are connected to the control unit with different cables. The outside noise sensor captures the environmental noise and transmits the information to the control unit. The control unit determines if the noise should be attenuated and contains a software algorithm which calculates the sound wave necessary to cancel out the environmental noise. If cancelation is appropriate, the control unit plays the corresponding sound wave through the speakers inside of the incubator, attenuating the noise via deconstructive interference. The residual noise sensor then captures the resulting noise inside of the incubator and communicates it to the control unit. Additionally, the device contains a voice pass through feature which allows parents or healthcare practitioners to communicate with the infant from the outside using the control unit and speakers.
This document outlines the acceptance criteria and study proving the device meets said criteria for the Neoasis active noise attenuation system.
Acceptance Criteria and Reported Device Performance
The primary acceptance criteria for the Neoasis device are based on its noise attenuation performance, especially within the specified frequency range, and its safety features. The direct acceptance criteria in the provided text are derived from the "Special Controls" section, which dictates the performance and safety requirements for this type of device.
| Acceptance Criteria Category | Specific Acceptance Criteria (from Special Controls) | Reported Device Performance (from "Performance Testing-Bench" and "Summary of Benefits") |
|---|---|---|
| (1) Non-clinical performance testing: Demonstrate device performs as intended under anticipated conditions of use. | (1)(i) Verification and validation of critical acoustic parameters, including the maximum output of the device. | |
| The device has a maximum output of 98 decibels. Software safeguards are in place to prevent malfunctioning and inappropriate noise production. The summary states that "software validation, verification, and hazard analysis was performed, and the results demonstrate that the software functions as intended, and the required specifications are met." The device outputs the calculated waveform through the speakers to cancel out environmental noise. | ||
| (1)(ii) Verification and validation of the attenuation performance of the device, including: | ||
| (A) Testing with compatible incubator model(s) and dimensions. | The device is only intended for use in the GE Healthcare Giraffe OmniBed Incubator. Performance testing was conducted simulating use with both devices on an infant mannequin within this incubator model. | |
| (B) Attenuation performance testing simulating different infant locations and orientations within the incubator. | Attenuation mapping across representative locations inside the incubator was performed. "Invictus Medical moved the noise sensing mannequin into different locations within the infant incubator to account for the potential movement of real infants." "The attenuation performance study described above evaluates the changes in performance within different locations of the incubator. This resulted in an attenuation map which demonstrates the device performance at different locations in the incubator." The results indicated the device performs as intended, with similar attenuation levels compared to the MiniMuffs. | |
| (C) Testing with relevant noise sources and room configurations. | "Invictus Medical used various sound sources that are present in a typical NICU setting" (e.g., Patient Monitor alarms, Ventilator alarms, Syringe Pump alarms, Male and Female Voices) to assess noise attenuation. The testing also evaluated situations where the infant could move. "Within the 250-1000 Hz range, the Neoasis and MiniMuffs device demonstrate similar attenuation performance." "Performance testing shows approximately a 7 decibels noise reduction, which is similar to MiniMuffs." | |
| (2) Software validation, verification, and hazard analysis. | Software validation, verification, and hazard analysis must be performed. | "Software verification and analysis was performed, and the results demonstrate that the software functions as intended, and the required specifications are met." The software has a moderate level of concern. |
| (3) Electrical safety and electromagnetic compatibility (EMC) testing. | Electrical safety and electromagnetic compatibility (EMC) testing must be performed for any electrical components of the device. | The device conforms to IEC 60601-1, IEC 60601-1-2, and IEC 60950-1 standards. |
| (4) Biocompatibility. | The patient- or user-contacting components of the device must be demonstrated to be biocompatible. | The device is classified as intact skin contacting with transient contact duration. Biocompatibility was evaluated in accordance with ISO 10993-1:2018 and FDA Guidance. Material information and manufacturing processes were provided in lieu of testing, supporting biological safety. |
| (5) Labeling. | Labeling must include specific information (instructions for infant placement, expected attenuation, warnings, cleaning/disinfection, compatible incubators). | Physician labeling includes indications for use, device description, warnings and precautions, expected benefit, and instructions for safe/effective use. It instructs users to clean/disinfect parts. Reusable components are cleaned/disinfected between patients. The labeling satisfies 21 CFR 801.109 and includes a summary of attenuation performance. Warnings regarding maximum output are implicitly covered by "warnings regarding the risks of exposure to the potential maximum output of the device". Instructions for infant placement and compatible incubators are also included. |
Specific Performance Requirements (from Indications for Use):
The attenuation performance of the device is for noises in the NICU with frequencies between 250 Hz to 1,000 Hz.
Reported Performance against Indication:
- "Within the 250-1000 Hz range, the Neoasis and MiniMuffs device demonstrate similar attenuation performance."
- Table 1 provides detailed attenuation data across multiple octave bands, including 250 Hz, 500 Hz, and 1000 Hz, for various noise sources. For many noise sources, particularly the Patient Monitor High (1000 Hz), Ventilator Medium (500 Hz), and Ventilator High (250 Hz, 500 Hz), Neoasis shows significant positive attenuation, often outperforming earmuffs. Even for other sources, where earmuffs might show slightly higher attenuation at certain specific frequencies, the Neoasis consistently demonstrates positive attenuation within or around the 250-1000 Hz range, indicating noise reduction. The summary confirms "approximately a 7 decibels noise reduction, which is similar to MiniMuffs."
Study Proving the Device Meets Acceptance Criteria
The study described is primarily a non-clinical bench testing study.
-
Sample size used for the test set and the data provenance:
- Test Set Description: The test set for the performance evaluation involved various simulated NICU noise sources and a mannequin with noise sensors.
- Sample Size: The sample size is not explicitly stated in terms of number of unique noise recordings or test runs, but the study evaluated the Neoasis device against multiple alarm types (Patient Monitor, Ventilator, Syringe Pump) from different brands (Philips, Maquet, Medfusion), at various alarm priorities (High, Medium, Low), and also included Male and Female Voices. For each noise source and alarm priority, attenuation was measured across 7 octave bands (125 Hz, 250 Hz, 500 Hz, 1000 Hz, 2000 Hz, 4000 Hz, 8000 Hz). Additionally, measurements were taken at "representative positions" within the incubator to account for infant movement, creating an "attenuation map."
- Data Provenance: The data was generated through prospective bench testing conducted by Invictus Medical. The location (country of origin) is not specified, but the sponsor information lists San Antonio, TX, USA.
-
Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
- For the bench testing performance study, no human experts were used to establish "ground truth" in the traditional sense of clinical label annotation. The ground truth here is the objective physical measurement of noise attenuation (reduction in decibels) using calibrated sensors (microphones) within the simulated environment. The "ground truth" is measured by the change in decibel levels of the noise before and after attenuation.
- For the Human Factors testing, 15 nurses were recruited to assess usability, but this was not for establishing "ground truth" for noise attenuation.
-
Adjudication method (e.g., 2+1, 3+1, none) for the test set:
- Given that the performance testing was a benchtop measurement study using sensors and objective physical parameters, there was no need for human adjudication. The measurements of sound levels are quantitative and derived directly from instrumentation.
-
If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance:
- No MRMC comparative effectiveness study was done. The device is an active noise attenuation system; it does not involve human "readers" or AI assistance in interpreting diagnostic images. Its purpose is to directly reduce noise. The comparison was device-to-device (Neoasis vs. MiniMuffs) based on physical properties (noise attenuation).
-
If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:
- Yes, a standalone performance evaluation was the primary method. The performance testing was of the Neoasis device acting independently to attenuate noise without human intervention during the "active" attenuation process. The device's software algorithm controls this process. Human factors testing was conducted to ensure safe and proper setup and use by a human, but the core function of noise attenuation is autonomous once setup.
-
The type of ground truth used (expert consensus, pathology, outcomes data, etc.):
- The ground truth for the performance test was objective physical measurement of sound levels (in decibels) before and after device operation, using a mannequin equipped with noise sensors (microphones). This is a direct measurement of the device's physical effect rather than a clinical outcome or human interpretation.
-
The sample size for the training set:
- The document does not specify a training set sample size for the device's internal software algorithm. The description states the control unit "contains a software algorithm which calculates the sound wave necessary to cancel out the environmental noise." This suggests a rule-based or real-time adaptive algorithm for active noise cancellation, rather than a machine learning algorithm that typically requires a discrete "training set" in the context of diagnostic AI. The software's function is verified through "software verification and analysis" and its performance through benchmark testing.
-
How the ground truth for the training set was established:
- As there's no mention of a traditional machine learning "training set" for the core noise cancellation algorithm, the concept of establishing ground truth for a training set does not apply directly to the information provided. The algorithm appears to be designed based on principles of active noise cancellation (deconstructive interference), and its "training" or optimization is likely integrated into its design and subsequent validation through performance testing against real-world noise inputs. The "ground truth" for its operation is the accurate generation of an anti-phase sound wave to reduce noise, which is validated by the measured attenuation levels during bench testing.
N/A